Political advertising has little persuasive power

Every four years, US presidential campaigns collectively spend billions of dollars flooding television screens across the country with political advertisements. But a new study co-authored by Yale political scientist Alexander Coppock shows that regardless of content, context or audience, these expensive commercials do little to convince voters.

The study, published September 2 in Science Advances magazine, measured the compelling impact of 49 high profile ads from the 2016 presidential campaign on a nationwide representative sample of 34,000 people through a series of 59 randomized experiments. The study builds on previous research suggesting that political advertisements have little impact on voter preferences, and shows that these weak effects are consistent regardless of a number of factors, including the tone of an advertisement, timing and the partiality of the audience.

“There’s an idea that a really good ad, or one that is delivered to a target audience in the right context, can influence voters, but we’ve found that political ads are consistently low in persuasion on a number of characteristics, ”said Coppock. Assistant Professor of Political Science at the Philosophical Faculty. “Positive ads don’t work any better than attack ads. Republicans, Democrats, and Independents respond similarly to ads. Ads broadcast in battlefield states are not much more effective than those broadcast in non-swing states. ”

Coppock and his co-authors – University of California-San Diego political scientist Seth J. Hill and UCLA political scientist Lynn Vavreck – conducted the study during the 2016 presidential primaries and general election.

Over 29 weeks, a representative sample of Americans were randomly divided into groups and assigned to watch campaign ads or a placebo ad – a commercial for auto insurance – before answering a brief survey.

The researchers selected ads based on real-time data on ad purchases and coverage of the top ads each week. They tested ads targeting or promoting Republican candidate Donald Trump and Democratic candidate Hillary Clinton, as well as commercials for primary candidates such as Republican Ted Cruz and Democrat Bernie Sanders. They analyzed the impact the ads had on respondents using multiple variables, including the candidate, party, or political action committee that sponsored them; whether they were positive or negative; the partiality of those who see the ads; the time until election day when they were broadcast; whether or not they were viewed in a battlefield state; and whether they aired during the primaries or general election.

They found that the ads, on average and across all variables, shifted respondents’ rating of a candidate’s favor by just 0.05 points on the survey’s five-point scale, which is small but given the size of the study is statistically significant, the researchers note. The effect of the ads on the people who wanted to vote was even less – a statistically insignificant 0.007 percentage point.

Campaigns should carefully consider whether advertising should be tailored to specific audiences, as the persuasiveness of advertising varies little from person to person or from advertisement to advertisement, the researchers conclude

The results do not prove that political advertising is always ineffective, said Coppock, noting that the study did not analyze the impact of an entire advertising campaign.

“TV commercials help candidates increase their public profile, which is extremely important, ”said Coppock, Resident Fellow at the Yale Institution for Social Policy Studies and the Center for the Study of American Politics. “Also, the effects we demonstrated were small but demonstrable and could mean the difference between winning and losing a close election.”

Andrew F. Carnegie Corporation, Marvin Hoffenberg Chair in American Politics and Public Policy at UCLA, and JG Geer, Dean of the College of Arts and Sciences at Vanderbilt University, supported the study.